"Doctoral Researcher in Language Technology"

We invite applications for a Doctoral Researcher in Language Technology for a fixed-term (3 years 8 months) employment contract from 1 January 2026, or as mutually agreed. We are seeking researchers with a background in Language Technology, Machine Learning, or Language Typology who are interested in automatic classification and analysis of hundreds of languages to undertake a PhD project exploring Unseen Languages in Language Identification. The project focuses on the automatic analysis of written texts, mostly already in digital format, while speech analysis remains entirely outside its scope.

The handling of unseen languages remains without real solutions, despite the research community having made a few practical excursions, such as using trash language models or confidence thresholds. This situation is partly due to the lack of proper investigation into the phenomenon and partly to its inherent complexity. In this project, we investigate the phenomenon of encountering previously unseen classes in text categorization, with a focus on language identification in texts. Well-defined and reliable open-set language identification methods will impact research fields that depend on resources created using language identification, such as text corpora for rare languages.

Language identification methods ground their predictions on training corpora for a finite number of separate languages. In practice, the size of this language repertoire has varied from two to well over a thousand languages. Almost exclusively, the methods label whatever text they encounter with a label of one of the languages in their repertoire. If they encounter text written in a language they have not seen before, they label it with what their algorithm deems the closest match. The results of such behavior can vary from the indicated language being a close relative to a seemingly random choice from the language repertoire. With more linear machine learning methods, it is possible to discern some of the reasoning behind the choices; however, this becomes more complicated with the newest deep learning-based systems.

To investigate the problem properly, we have assembled a highly skilled group of international collaborators with whom we will examine several case studies where unseen languages present practical challenges for researchers or the users of language resources created by them. Using existing text collections, we will create a high-quality evaluation setting for open-set LI. Using it, we further adapt and develop the current state-of-the-art methods we have devised for handling unseen languages.

The exact role and responsibilities of the doctoral researcher within the project will be determined by their background and research interests.

Requirements for the position and assessment of qualifications

According to the Regulations of the University of Helsinki, an appointee to the position of doctoral researcher shall hold a second-cycle degree and the right to complete a doctoral degree at the University of Helsinki or shall be granted the right to complete a doctoral degree at the University of Helsinki during the trial period at the beginning of the employment relationship. Other requirements shall include an approved research proposal, as well as the ability and motivation, as demonstrated in previous degree studies and otherwise, to pursue postgraduate studies and a doctoral degree according to the study plan and research proposal. When assessing applicants’ qualifications, attention will be paid to the applicant’s previous studies and knowledge relevant to the goals of the Unseen Languages in Language Identification project. Evidence of independent and collaborative research is an advantage. We look forward to receiving applications from students with diverse backgrounds. From applicants with a background in linguistics, we expect a strong interest in machine learning methods and algorithms. From applicants with a background in machine learning, we expect a strong interest in the analysis of languages and language typology.

What we offer

The selected candidate will receive a full salary for 3 years and 8 months and will be assigned two doctoral supervisors. The salary for the position will be based on levels 2-4 of the job requirement scheme for teaching and research personnel in the salary system of Finnish universities. Additionally, the appointee will receive a salary component based on their personal performance. The annual gross salary ranges from €31,000 to €39,000. The position is filled with a six-month trial period. In Finnish universities, the teaching and research staff have a special system regarding the free time equivalent to holidays. They belong to the 1,612 annual working hour system, in which the annual amount of paid holiday is app. 5-6 weeks on average per year. The University of Helsinki offers comprehensive services to its employees, including occupational health care and health insurance, sports facilities, and opportunities for professional development. The University provides support for internationally recruited employees with their transition to work and life in Finland. The selected candidate will also have access to travel money and limited funds for open-access publishing.

How to apply

The application must be accompanied by the following documents in PDF format: A curriculum vitae; Names of two referees; A preliminary research plan for the doctoral research, consisting of a brief description of previous research experience and a description of the planned research; Study transcript of the master’s degree; Publication’s list. Applications must be submitted through the University of Helsinki electronic recruitment system.